Pinned Tweet
Replying to @RandyWKirk1
I’m calling it now. $spcx will be trading at $500+ by end of FY28 based on 10GW of net new data center lit up. This trend will continue thru 2028 with addtional 10GW. During 2028 orbital compute will start coming online. I expect orbital to exceed terrestrial by 2029. By 2030 orbital compute will dominate and net new terrestrial will scale back. The economics of orbital compute win over terrestrial and that is the determining factor. @elonmusk called it early. Unlimited power, no need for batteries, generators or cooling. It is first principle compute being executed.
4
241
poshimoto retweeted
Antonio Gracias has invested with Elon since PayPal. He sat on Tesla's board through the Model 3 crisis.
He says one night explains Elon better than anything.
It's 2018, 3 AM, in the body shop. By Elon's own account, Tesla was about a month from bankruptcy.
Elon sits down with the engineers and starts cutting welds out of the car. About 30. Math in his head.
"If he had been wrong, we'd have gone bankrupt."
The simulation came back. He was right.
"It was like watching Michael Jordan play basketball times 10."
Gavin Baker, sitting next to him, on why Elon started xAI: so AI isn't "just dominated by OpenAI, Google."
That panel was two years ago. xAI was a startup with no product.Gavin Baker explained why Elon started xAI before most people took it seriously.
"To make sure that it is not just dominated by OpenAI, Google."
Then Antonio Gracias, who sat on Tesla's board, explained why Elon can actually pull it off.
It's 2018, 3 AM. The Model 3 body shop is failing. Tesla is running out of money.
"He was running the math in his head… take about 30 welds out of the car."
"If he had been wrong, we'd have gone bankrupt."
He wasn't wrong.
"It was like watching Michael Jordan play basketball times 10."
That's the same brain now running xAI. Back then it was 30 welds. Now it's a million GPUs.
Everyone asks if xAI can catch OpenAI and Google. The people who've seen Elon at 3 AM don't ask.
Today it's inside SpaceX, running one of the largest GPU clusters on Earth.
Both of them called it early. Both had seen 3 AM.
poshimoto retweeted
Made an edit to recap @SpaceX's historic 15th flight of Starship, which included the first-ever catch of the ship.
🚀
poshimoto retweeted
This is f*cking insane. SpaceX just caught Starship with the tower. Elon Musk and Gwynne Shotwell sat down to explain what comes next:
"If one agrees that AI is a potential risk to the public, there should be some referee making sure companies don't cut corners"
In this 1-hour talk, they cover AI risks, Starship, Terafab and how SpaceX is actually run
Grok Bot → SpaceX Engineering → Starship → Terafab → A Company That Runs on AI
Most people saw a rocket land. This is the hour about the machine behind it
Worth more than any $2,000 course
Bookmark and watch it today
Then read how one person runs 6 Grok bots and 200+ agents ↓
SpaceX caught Starship, so what now?
Before SpaceX can land astronauts on the Moon, it has to refuel a Starship in orbit. That could take 6 tanker launches. Or more than 20.
Here's how the whole mission works.
Starship reaches orbit almost empty, so a depot ship waits up there and tankers fill it one by one. Each tanker docks alongside, small thrusters settle the propellant at the back of the tanks, and tank pressure pushes it across.
Musk says 6 to 8 tankers and NASA officials agree. A NASA safety adviser said about 22.
So far propellant has only been moved between tanks inside one ship, on Flight 3 in March 2024. The first ship-to-ship test is planned for early 2027.
Then four astronauts launch in Orion and dock to the lander's nose in Earth orbit. Starship fires and pushes both of them to the Moon, about four days away.
In lunar orbit two of the crew move into Starship and head for the south pole. Raptors do most of the braking, then thrusters high on the hull can take over for the last 100 metres so less dust gets thrown up. A few days on the surface, back up to Orion, and Orion brings all four home.
And those big wings on the lander? Solar arrays, about 18 metres each going by SpaceX's renders. No heat shield and no flaps, because it never comes back.
The plan is a docking test in 2027 and a first landing in 2028. NASA hasn't picked the lander yet, and these dates have slipped before.
Still to prove: ship-to-ship refuelling, keeping propellant cold for weeks, launching about once a week, and landing on the Moon without a crew.
This week Starship reached orbit again and the tower caught the ship for the first time. Now it needs to do that every week.
Sources: NASA, SpaceX, ESA/Airbus, NASASpaceflight, SpacePolicyOnline, AmericaSpace
Adam Jonas from $MS about the success of $SPCX Starship 15
“In our view, Flight 15 is a historic inflection moment for the company and spaceflight as a whole. SpaceX has recovered three Starship + many hundreds of Falcon 9 boosters in its history, and now others have accomplished the same. However, no entity has ever even attempted to recover a rocket upper stage in history (not counting Space Shuttle) given the intense thermal and mechanical stress of orbital re-entry.
With an upper stage now recovered intact, SpaceX can conduct a deep inspection of Ship 42 and use the findings to improve future vehicles in a way that was never previously possible. We believe it also helps clear the path to ramping full reuse of both booster and ship, the last major technical milestone before weekly or even daily Starship flights become realistic.
With a ship catch, SpaceX has materially de-risked the longer-term TAMs across all 3 of its segments.”
poshimoto retweeted
My current thoughts on what happened to Jensen.
He was fighting three big policy disputes.
Externally, he was lobbying to allow some degree of chip sales to China and maybe even willfully turning a blind eye to some restricted sales “leakage” into China. US China hawks in both political parties were fighting him. Oddly, Beijing appears to be in agreement with the China hawks since they would rather suffer the short term pain of limited access to frontier chips to force domestic innovation of their own AI ecosystem.
I think Jensen had the correct take on this issue. Having China dependent on US technology is far better for the US than China developing its own AI tech.
Internally, he had a vendor financing policy to lock in customers not only to his chips, but to his financing. The CFO apparently didn’t agree. Maybe that’s why Wall Street became skittish on vendor financing - would be interesting to know what she was saying to them.
Again, I think Jensen was also correct here.
Strategically, he embraced open source to undercut Anthropic & OpenAI dominance. Needless to say, they didn’t appreciate this and may have done something about it.
Again, Jensen was right.
None of us know what went down last week. We don’t even know if he resigned in shame, resigned because of possible legal pressures from the above policy disputes, or if the Board asked him to resign because of the same.
Given he has hired high powered lawyers, we aren’t likely to hear from Jensen anytime soon either. We will find out when one of the above factions threatens him legally, either via civil suit, criminal probe, or Congressional action.
Maybe I’m in a bubble, but I think Jensen doesn’t realize how much support he has among rank and file shareholders, company employees, and the AI industry at large.
Not sure if any of that matters given what may, or may not, unfold in the coming weeks.
What I am pretty sure of though is that Nvidia as a company has been severely wounded. Its supply chain partners will no longer accept projections based on the CEOs word. Customers can no longer rely on vendor financing, nor even on massive outsized supply given its supply chain relationships have been broken. Engineers are already starting to jump ship. Competitors are now emboldened, their stock prices climbing as Nvidia’s falls.
poshimoto retweeted
马斯克抛出一个相当直接的条件:任何英伟达员工,只要黄仁勋没能回到自己的位置上,就可以辞职去 SpaceX 和芯片项目 Terafab。
这话表面是招人,实际是冲着英伟达董事会去的。潜台词很清楚:如果你们让公司的核心人物离开,那全球最稀缺的一批 GPU 工程师,门就在这里。
对英伟达来说,失去黄仁勋和这批工程师,剩下的可能只是一个没有灵魂的名字。对马斯克来说,这是一次把顶尖人才直接装进自己芯片计划的机会。
压力能不能逼董事会回心转意,还是说我们真要看到科技史上最大规模的一次工程人才外流,现在还不好说。
poshimoto retweeted
I have spent over six years of my life working toward this moment.
Thousands of hours. Countless failures. Lessons learned. Sacrifices made. An extraordinary team that refused to accept the impossible.
Today, I watched a Starship return from space and get caught by the very tower that launched it.
And knowing that the heat shield we poured our hearts into helped bring her home is a feeling I will carry for the rest of my life.
To every person who ever turned a wrench, cast a tile, solved a problem, or refused to give up: WE DID IT.
We didn’t just catch a spaceship today.
We caught a glimpse of humanity’s future!
Welcome home S42!!!
This is a bigger deal than most will comprehend. This validates the entire business proposition of reusable orbital vehicles. This will drive the cost to orbit significantly. Everything that has been put into orbit previously was done based on the cost of the vehicle being discarded at the end of the mission. Very expensive. Imagine how much it would cost to fly if they only could use an airplane once. Orbital compute is now a reality and will start in late 2027 and scale in 2028. It’s go time $spcx investors! Congrats to @elonmusk and the entire @SpaceX Starship team.
2
27
4D Chess. Checkmate
马斯克这次可能真要捡个大便宜了。🤣
最新消息,特斯拉与SpaceX合并谈判取得进展,最快12月公布协议。
老黄刚离开英伟达,马斯克就表示,能雇佣黄仁勋是他的荣幸。
我突然有个大胆的想法:
假如两家公司合并,再把老黄请过去。特斯拉负责机器人和自动驾驶,SpaceX负责卫星、算力和AI基础设施,xAI负责大模型,老黄负责芯片和融资。
别忘了,黄仁勋以前不光卖GPU,还给AI客户提供融资,活脱脱一个AI央行行长。
现在英伟达的新掌门人想收紧钱袋子,老黄要是跑到SpaceX继续放水……
以后AI公司缺钱,可能不用找英伟达了,直接去找马斯克申请贷款。🤣
一个能造火箭、造芯片、造机器人的AI央行,会把华尔街折腾成什么样。😃😃😃
#SpaceX #TSLA #NVDA #AI
19
poshimoto retweeted
RUMOR: Elon Musk has promised jobs at SpaceX and Terafab to any Nvidia staffer who leaves if Jensen Huang isn’t reinstated.
Readers added context they thought people might want to know
The rumor originates from a fabricated Nvidia employee letter copied from the 2023 OpenAI letter; Nvidia VP Bryan Catanzaro called it obviously fake with no supporting evidence, and Musk has only said he would be honored to hire Huang personally or sign a petition.
x.com/ctnzr/status/2…
talkesport.com/business/nvidi…
x.com/elonmusk/statu…
poshimoto retweeted
现在基本可以确定,Anthropic就是这次针对黄仁勋抹黑行动的幕后黑手。向纽约时报爆料老黄和Alice私事的,正是Anthropic联合创始人Jack Clark。而他手里持有的Anthropic股票比Dario还多200万股。除了我之前分析过的双方在对华芯片出口问题上的分歧之外,还有一个几乎没人讨论过的关键问题。
英伟达其实早就准备亲自下场,跟Anthropic和OpenAI抢饭碗了。今年6月,英伟达宣布了Nemotron 3 Ultra,一个拥有5500亿参数的开放权重模型。老黄在四个月前的这段采访里甚至亲口说,Nemotron已经可以跟Claude Opus相提并论,具备接近前沿大模型的能力,而且成本还便宜得多。
你站在Anthropic的角度想想,如果这时候还坐以待毙,那不就是等死吗?一方面,你整个公司的命脉都掌握在英伟达手里,因为训练和运行模型都离不开它的GPU。另一方面,既然英伟达自己掌握着AI模型最核心的瓶颈——算力,那它凭什么不自己训练大模型,直接跟你竞争?AI产业就像一个五层蛋糕,而英伟达完全有潜力把这五层全部吃下来。
我从来不相信什么巧合,而且我觉得很明显,Dario这次就是在故意陷害黄仁勋。老黄在美国和中国的科技圈、商界、政界,几乎跟所有人关系都很好。唯一跟他过不去的,就是Dario。
老黄一直认为,开源模型对AI行业的发展是有好处的,而让中国的AI模型跑在英伟达的芯片上,对美国也是有利的。过去几年,中国出了不少开源大模型,最有代表性的就是Kimi K3。也正是在这个时候,老黄专门注册了一个X账号,就是为了替开源AI说话。他是真的相信开源这条路,而且这对英伟达的生意本身也有好处。
Dario则一直是开源AI最主要的反对者,还多次公开批评老黄向中国出售芯片。大家要明白,Anthropic本身就高度依赖英伟达的芯片。换成一个正常人,就算心里不同意老黄的观点,为了生意也会选择闭嘴。但Dario却一次又一次公开跟老黄唱反调,甚至不惜冒着得罪英伟达的风险。这说明他在这件事上的立场有多坚决。
为什么?因为Dario自己也很清楚,对于绝大多数任务来说,大部分企业和个人根本没必要花高价使用Claude,直接用便宜的中国开源模型就够了。要阻止这种情况,唯一的办法就是让美国政府出手,要么禁止中国AI模型,要么禁止英伟达向中国出售芯片。
但第一个办法有个很大的问题:一个在互联网上随便就能下载的软件,你怎么禁止别人使用?第二个办法的问题则是,老黄跟共和党的关系非常好,尤其是川普。偏偏川普又特别不喜欢Dario。Dario是美国科技巨头老板里少数偏向民主党的人之一,川普甚至还禁止美国政府使用Claude。
所以过去两年,Dario可以说是有气没地儿发。我怀疑他就是一直在等,等到民主党赢下中期选举之后,才终于对老黄发动了这次攻击。大家别忘了,Alice Huang这件事根本不是什么新闻。中文推特上的人早在一年前就已经知道了,川普估计也早就知道了。
你们真的相信这是巧合吗?为什么这件事偏偏在美国中期选举刚结束、距离Anthropic上市只剩几个星期的时候,突然被爆了出来?
poshimoto retweeted
Let’s make Jensen an offer he can’t refuse, portfolio includes:
1. AI Chip Design: AI6, AI7, AI8 and beyond.
2. Terafab: Advanced semiconductor manufacturing at scale.
3. Orbital AI Infrastructure: Design and scale space-based AI data centers.
4. Physical AI: Optimus humanoid robots and next-generation autonomous products.
poshimoto retweeted
The most valuable thing in life isn’t money, it’s true friendship.
"A friend is someone who's going to support you in difficult times. I suppose a friend in need is a friend indeed, like if someone's still supporting you when the chips are down, that's a friend."
— @elonmusk
poshimoto retweeted
Had a little time on my hands this morning to do some writing… send along to someone you know who may benefit
---
“Practical principles to counter a modern-day coup/psyop/struggle session”
1) Own what you did but fiercely resist accommodating the mob, resist owning half-truths or actions/behaviors you didn’t do.
2) The op is not about your past actions, it’s about your enemies and this event as their opportunity
3) No matter how bad it gets, nothing is as bad as it seems.
4) Find the enemies at the very root of the op and find every creative problem-solving tool at your disposal to neutralize them.
5) People will throw their grandmother under the bus if they believe doing so will protect their reputation. They can become highly useful idiots in the op. Their power and access must be fiercely and quickly cut off.
6) Do not hesitate in cutting off anyone you suspect of disloyalty.
7) Build a team of close and loyal patriots and have one of them run a dedicated, well-resourced war room to put the rebellion to rest.
8) Use your patriots and regular meditation to buoy your psychological resistance to the op.
9) Your enemies aren’t as smart or hard working as you. Outsmart and outwork them as a way of life until the op is over.
10) Think about the op as if it were in the past. What are the difficult things you procrastinated on or avoided that would have stopped the op? Do them.
11) Never be ashamed of the op. Business as usual. Continue on with optimism and joy in your everyday work and its mission.
12) Download telegram and use it as your messaging application for all sensitive communications.































